List your company

Long-context Reasoning Accuracy on LongBench

70.4Accuracy (LongBench)CAP-CoT
51.99256.77161.5566.329Mar 13, 2026Mar 20, 2026Mar 27, 2026Apr 3, 2026Apr 10, 2026Apr 17, 2026Apr 25, 2026

Evaluation results

MethodLinks
CAP-CoTBackbone=GPT-4o2026.0470.4
CAP-CoTBackbone=Deepseek-V32026.0469.5
AoTBackbone=GPT-4o2026.0469.5
CAP-CoTBackbone=GPT-4o-mini2026.0469.3
CAP-CoTBackbone=Qwen-turbo2026.0468.9
AoTBackbone=DeepSeek-V32026.0368.7
AoTBackbone=Deepseek-V32026.0468.7
AoTBackbone=GPT-4o-mini2026.0368.5
AoTBackbone=GPT-4o-mini2026.0468.5
AoTBackbone=Qwen-Turbo2026.0368.1
AoTBackbone=Qwen-turbo2026.0468.1
ECONBackbone=GPT-4o2026.0466
ECONBackbone=Deepseek-V32026.0465.2
GoTBackbone=GPT-4o2026.0465.2
MADBackbone=GPT-4o2026.0465.2
ECONBackbone=GPT-4o-mini2026.0465
MADBackbone=Deepseek-V32026.0464.5
ToTBackbone=GPT-4o2026.0464.5
MADBackbone=GPT-4o-mini2026.0464.2
ECONBackbone=Qwen-turbo2026.0464
GoTBackbone=GPT-4o-mini2026.0463.8
AFlowBackbone=GPT-4o2026.0463.5
GoTBackbone=DeepSeek-V32026.0363.4
GoTBackbone=Deepseek-V32026.0463.4
ToTBackbone=DeepSeek-V32026.0363.2
MADBackbone=Qwen-turbo2026.0463.2
ToTBackbone=Deepseek-V32026.0463.2
GoTBackbone=GPT-4o-mini2026.0363.1
ToTBackbone=GPT-4o-mini2026.0463.1
ToTBackbone=GPT-4o-mini2026.0362.8
GoTBackbone=Qwen-Turbo2026.0362.8
GoTBackbone=Qwen-turbo2026.0462.8
ToTBackbone=Qwen-Turbo2026.0362.4
ToTBackbone=Qwen-turbo2026.0462.4
FoTBackbone=GPT-4o2026.0461.5
AFlowBackbone=DeepSeek-V32026.0361.4
AFlowBackbone=Deepseek-V32026.0461.4
AFlowBackbone=GPT-4o-mini2026.0361
AFlowBackbone=GPT-4o-mini2026.0461
PromptAgentBackbone=GPT-4o2026.0461
CoT-SCBackbone=GPT-4o2026.0460.8
CCoTBackbone=GPT-4o2026.0460.5
AFlowBackbone=Qwen-Turbo2026.0360.4
AFlowBackbone=Qwen-turbo2026.0460.4
Self-RefineBackbone=GPT-4o2026.0460.2
TDA-RCBase LLM=DeepSeek-V32026.0359.5
TDA-RCBackbone=DeepSeek-V32026.0359.5
CoTBackbone=GPT-4o2026.0459.5
TDA-RCBase LLM=GPT-4o-mini2026.0359.3
TDA-RCBackbone=GPT-4o-mini2026.0359.3
FoT (n=8)Backbone=DeepSeek-V32026.0359.3
FoTBackbone=Deepseek-V32026.0459.3
PromptAgentBackbone=Deepseek-V32026.0459.2
FoT (n=8)Backbone=GPT-4o-mini2026.0359.1
FoTBackbone=GPT-4o-mini2026.0459.1
TDA-RCBase LLM=Qwen-Turbo2026.0359
Instruction InductionBase LLM=DeepSeek-V32026.0359
FoT (n=8)Backbone=Qwen-Turbo2026.0359
TDA-RCBackbone=Qwen-Turbo2026.0359
FoTBackbone=Qwen-turbo2026.0459
CoT-SCBackbone=DeepSeek-V32026.0358.9
CoT-SCBackbone=Deepseek-V32026.0458.9
Instruction InductionBase LLM=GPT-4o-mini2026.0358.8
CCoTBackbone=GPT-4o-mini2026.0458.8
Role / Persona PromptingBase LLM=DeepSeek-V32026.0358.7
PromptAgentBackbone=GPT-4o-mini2026.0458.7
Role / Persona PromptingBase LLM=GPT-4o-mini2026.0358.6
HoTBase LLM=DeepSeek-V32026.0358.6
CoT-SC (n=5)Backbone=GPT-4o-mini2026.0358.6
CoT-SCBackbone=GPT-4o-mini2026.0458.6
Prompt CanvasBase LLM=GPT-4o-mini2026.0358.5
Prompt CanvasBase LLM=DeepSeek-V32026.0358.5
PromptAgentBackbone=Qwen-turbo2026.0458.5
CCoTBackbone=Qwen-turbo2026.0458.5
CCoTBackbone=Deepseek-V32026.0458.5
HoTBase LLM=GPT-4o-mini2026.0358.4
CoT-SC (n=5)Backbone=Qwen-Turbo2026.0358.4
Self-RefineBackbone=DeepSeek-V32026.0358.4
CoT-SCBackbone=Qwen-turbo2026.0458.4
Self-RefineBackbone=Deepseek-V32026.0458.4
Instruction InductionBase LLM=Qwen-Turbo2026.0358.3
HoTBase LLM=Qwen-Turbo2026.0358.2
Self-RefineBackbone=GPT-4o-mini2026.0358.2
Self-RefineBackbone=GPT-4o-mini2026.0458.2
Role / Persona PromptingBase LLM=Qwen-Turbo2026.0358.1
Prompt CanvasBase LLM=Qwen-Turbo2026.0358
Self-RefineBackbone=Qwen-Turbo2026.0358
Self-RefineBackbone=Qwen-turbo2026.0458
CoTBackbone=DeepSeek-V32026.0357.7
CoTBackbone=Deepseek-V32026.0457.7
CoTBackbone=GPT-4o-mini2026.0357.6
CoTBackbone=GPT-4o-mini2026.0457.6
CoTBackbone=Qwen-Turbo2026.0357.3
CoTBackbone=Qwen-turbo2026.0457.3
APBackbone=GPT-4o2026.0455.2
Analogical PromptingBase LLM=DeepSeek-V32026.0353.4
APBackbone=Deepseek-V32026.0453.4
Analogical PromptingBase LLM=GPT-4o-mini2026.0352.9
APBackbone=GPT-4o-mini2026.0452.9
Analogical PromptingBase LLM=Qwen-Turbo2026.0352.7